AI-powered MRI analysis offers automated risk stratification via body composition in multiple myeloma patients.
Key Details
- 1A deep-learning system was trained and validated on 39 whole-body MRI scans in multiple myeloma patients.
- 2The pipeline enables automated tissue and organ segmentation in seconds or minutes, speeding up analysis and reducing operator time.
- 3Longitudinal MRI analysis revealed that decreases in skeletal muscle and increases in fatty tissue during treatment are statistically significant (p < 0.001).
- 4Higher baseline subcutaneous adipose tissue predicted better progression-free survival (HR = 0.60–0.67), while increases in visceral fat correlated with worse survival (HR = 2.89).
- 5Including body composition metrics in routine MRI adds no burden but may facilitate earlier interventions and better patient outcomes.
Why It Matters
This work highlights how AI can transform standard imaging into a biomarker-rich tool that informs prognosis and individualized care in oncology, maximizing existing imaging data without added patient risk or discomfort.
Source
AuntMinnie
Related News
•Radiology Business
AI Targets Radiology Residents’ Training Gaps for Personalized Learning
Artificial intelligence can identify gaps in radiology residents' clinical exposure, supporting tailored educational interventions.
•Radiology Business
AI Work List Triage Tool Dramatically Reduces Radiology Turnaround Times
AI-based work list prioritization in radiology slashes reporting and turnaround times by up to 90%.
•AuntMinnie
AI Tool Cuts Radiologist Reporting Time for Chest CTs by 15%
Real-world study finds commercial AI for pulmonary nodule detection reduces chest CT reporting times by nearly 15%.